基于案例推理的地铁非常规突发事件应急决策方法研究  

Research on emergency decision method of unconventional emergency in subway based on case-based reasoning

作　　者：徐亚博[1] 汪彤[1] 王培怡[1] 宋冰雪[1] 葛悦[1] 

机构地区：[1]北京市劳动保护科学研究所,北京100054

出　　处：《中国安全生产科学技术》2013年第8期44-48,共5页Journal of Safety Science and Technology

基　　金：北京市自然科学基金资助项目(9123027);灭火救援技术公安部重点实验室开放课题(KF2012001)

摘　　要：针对地铁非常规突发事故发生时,由于地下环境复杂且封闭,很难获取较全面、准确的事件信息等问题,非常规突发事件的地铁应急决策是一个复杂的过程。通过对地铁事件原则性机理进行分析和研究,得到地铁发生重大火灾事故情景分析的贝叶斯网络节点变量,借鉴CBR模型的思想,将基于结构和属性的双重情景检索的事故案例推理应急决策方法应用到地铁火灾非常规突发事件应急决策中,该方法排除了因属性值缺失造成相似度无法计算或计算有误的情况,可为地铁交通应急平台的后期建设提供理论支持。Due to the complex and closed environment in underground subway system, it is very difficult to obtain more comprehensive and accurate accident information when a major unconventional accident occurs. So, emergen- cy decision-making of subway unconventional event is a complicated process. Through analysis and research of the subway event mechanism principle, Bayesian networks structure of subway fire accident scenario was get. Refer- ence to the ideas of the CBR model, emergency decision method based on the structure and properties of dual scene retrieve is applied to the subway fire accident case. The case study results show that the method can overcome the drawbacks of inaccuracy in similarity calculation due to lack of attribute values, avoid the invalidation of retrieval system caused by information loss, and provides theoretical support for subway traffic emergency platform construc- tion.

关 键 词：案例推理 非常规突发事件 情景检索 应急决策 

分 类 号：X928.03[环境科学与工程—安全科学]